UC San Diego

3D genome structure plays an important role in gene regulation. However, the study of genome folding in the context of cell-type specificity is largely missing. Here, we developed a single-cell multiomic assay, snm3C-seq, to simultaneously profile chromosome architecture and DNA methylation in the same single cell, enabling the investigation of 3D genome contacts across a wide range of cell types in a complex tissue. Nevertheless, it remains challenging to identify 3D chromatin features and compare them between cell types with single-cell data, given the heterogeneity of genome structures across cells as well as the limited reads being detected within a cell. To handle this question, we developed scHiCluster, a comprehensive framework for single-cell chromosome conformation analysis. The framework calculates low-dimensional embedding of single cells for accurate cell type clustering based on chromatin contacts, followed by annotating cell type specific chromatin loops and domains. We applied snm3C-seq and scHiCluster to developing human brain samples to analyze 53,063 cells. We identified 139 cell populations organized into 10 major lineages in the datasets, as well as over 2.5 million putative regulatory elements from these populations. The regulatory elements at chromatin loop anchors are highly enriched for putative causal common genetic variations of schizophrenia. We also observed difference in developmental timing between DNA methylation and 3D genome structures, and chromatin conformation changes often prime CG methylation (mCG). Together, this thesis shows the development of experimental and computational workflows to investigate single cell 3D genome structures and how we apply the technologies to study the molecular dynamics during brain development.